• Publications
  • Influence
Learning Transferable Architectures for Scalable Image Recognition
TLDR
We study a method to learn the model architectures directly on the dataset of interest and apply them to ImageNet. Expand
  • 1,959
  • 413
  • PDF
Neural Architecture Search with Reinforcement Learning
TLDR
We use a recurrent network to generate the model descriptions of neural networks and train it with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set. Expand
  • 2,083
  • 301
  • PDF
Efficient Neural Architecture Search via Parameter Sharing
TLDR
We propose Efficient Neural Architecture Search (ENAS), a fast and inexpensive approach for automatic model design. Expand
  • 952
  • 228
  • PDF
Progressive Neural Architecture Search
TLDR
We propose a new method for learning the structure of convolutional neural networks (CNNs) that is more efficient than recent state-of-the-art methods based on reinforcement learning and evolutionary algorithms. Expand
  • 784
  • 147
  • PDF
SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
TLDR
We present SpecAugment, a simple data augmentation method for speech recognition that achieves state-of-the-art performance on the LibriSpeech 960h and Swichboard 300h tasks, outperforming all prior work. Expand
  • 510
  • 101
  • PDF
AutoAugment: Learning Augmentation Policies from Data
TLDR
We use a search algorithm to find the best policy such that the neural network yields the highest validation accuracy on a target dataset. Expand
  • 470
  • 83
  • PDF
Searching for Activation Functions
TLDR
We use a combination of exhaustive and reinforcement learning-based search techniques to discover novel activation functions. Expand
  • 668
  • 80
Transfer Learning for Low-Resource Neural Machine Translation
TLDR
We present a transfer learning method that significantly improves Bleu scores across a range of low-resource languages. Expand
  • 353
  • 58
  • PDF
Understanding and Simplifying One-Shot Architecture Search
TLDR
We show that it is possible to efficiently identify promising architectures from a complex search space without either hypernetworks or reinforcement learning controllers. Expand
  • 256
  • 55
  • PDF
AutoAugment: Learning Augmentation Strategies From Data
TLDR
We use a search algorithm to find the best policy such that the neural network yields the highest validation accuracy on a target dataset. Expand
  • 313
  • 49
  • PDF